A Bayesian analysis of a zombie drug trial finds a 6% to 39% cure rate and 93% odds the drug beats existing treatments. The instructor shows how the same principles Alan Turing used to crack Enigma code power this simple statistical model.
Chapters
Great. You've done some Bayesian data analysis and you've summarized the result. Here are the code and the output you produced in the last exercise. If we would take this result, write it up and send it to the International Journal on Zombiology, it could go something like Given the data of 2 cured and 11 relapsed zombies and using the Bayesian model described before, there is a 90% probability that our drug cures between 6% and 39% of treated zombies. Further, there is a 93% probability that our drug cures zombies at a higher rate than the current state of the art drugs. Sounds like science to me. Now this was a very simple Bayesian model. There was just one parameter to estimate the underlying proportion of success, but it uses the exact same principles and machinery as the Bayesian model Alan Turing used to crack the Enigma code. But what is that machinery? In the next chapter we're going to open up the box and take a deeper.